Tim Petri

I have received three separate notifications about College Basketball from @coinbase in the past *hour* alone. It is absurd that, amidst arguably the worst collapse in trust in this industry’s history, the largest American CEX has completely pivoted to trying to get their customer base hooked on sports gambling, so that they can extract even more exorbitant fees. At this point, it is undeniable that Coinbase *is* part of the industry’s problem. I will be ending my Coinbase One subscription and moving my business to new a CEX, any recommendations?

@zenorocha Are you guys still using inngest? Or did you quickly end up moving to internal orchestration?

Resend is profitable. It's been the case since last year, but it happened by accident. When we started, we knew we were going after giants. Companies with thousands of employees, hundreds of millions raised, some already public. So we took the VC route. Got into YC, raised some money. Then one day, I looked at the projections and realized we were close. We didn't try to cut our cloud bill, reduce our DataDog costs, or slow down hiring to get there earlier. No, we just kept focused on building the product, improving the infrastructure, and making sure customers are happy. Everyone told us to burn cash and grow faster. The reality is, you can grow fast, be aggressive, and still be profitable. People mistake profitability with being anti-VC. The right investors can be incredibly helpful (that was the case for us at least). Being profitable means optionality. You can raise more or maybe less. You can raise now, or maybe later. Having that choice is better than not having it. When you're profitable, every decision is about the product, not the runway.

Thinking about starting another side project.

@BernieSanders How exactly do imagine this "fight" to go down? This technology is inevitable, and you politicians are not doing anything to address the impending worker surplus.

Jeff Bezos, worth $234 billion, plans to replace 600,000 Amazon workers with robots. Now, he wants to spend $100 billion to fully automate not just his warehouses, but factories in the U.S & other countries. Oligarchs are waging all out war against workers. FIGHT BACK.

@typesfast It was @JeffDean

Whoever named Gemini at Google really named it. In the mythology, the immortal twin gives up his immortality to save the life of his mortal twin. It’s just like Google giving up 100% of its free cash flow to make sure Deepmind survives.

Prediction: This will turn into the greatest security Whack-a-Mole ever. Probably before any real physical embodiment of these systems is deployed at scale.

@skirano How does this work if designs are built using a design system? Components for that design system are already in repo, with a11y, etc. Likely shadcn based.

Introducing Figma Connect. The best way to turn your Figma designs into code. No MCP hell. No plugins. Just copy and paste your designs into MagicPath and turn them into interactive prototypes without compromising your craft. Every pixel. Every detail. Every asset. Preserved.

@dvoytenko @vercel @cramforce @rauchg Mixed up two parameters, but for posterity: We set traceSampleRate to 100 as well.

@dvoytenko @vercel @cramforce @rauchg Hey Dmytro. Is there a specific version I must use for this? Do you have an example of configuring sampling? FWIW right now we don't touch sampling at all on the backend. We set sessionSampleRate to 100 in datadog rum.

Am I missing something or should the @vercel opentelemetry library not accept and propagate tracecontext coming from the frontend for the purposes of distributed tracing? @cramforce @rauchg

Today, Arch is happy to announce the closing of a $75M tokenized CLO with @galaxyhq issued on @avax, to continue to finance our crypto-backed loans. This is institutional credit infrastructure on modern rails. Tokenization is finally moving from theory to production.

Thanks for replying! This I already know, and I've been able to successfully propagate context to further downstream services. My issue is around incoming request propagation in nextjs. The docs claim it works but that's what I'm observing. In our case, DatadogRum attaches tracecontext headers to all requests client-side but the default spans created by the nextjs framework (see docs [0] for list, but canonical would be the [http.method] [next.route]) fail to make use of the incoming trace context. [0] #httpmethod-nextroute" target="_blank" rel="nofollow noopener">nextjs.org/docs/app/guide…

@timpetri @vercel @cramforce @rauchg The trace context propagation via fetch and http(s) only happen for the same origin requests by default. That's because sometimes trace context had sensitive data. But you can expand via `fetch` instrumentation config

@vercel_dev @vercel_support

It claims to come pre-configured with tracecontext propagation for both inbound and outbound but the server spans created by the nextjs framework ignores any trace context. The result is you have orphan SERVER HTTP spans.

@DrizzleORM Oh how I wish I could rip out prisma from our codebase. I don't understand how anyone who has used prisma/accelerate would dare to try their postgres offering.

- Drizzle drops benchmarks, running http requests on server+db - prisma blames us for using search in ecom app and they don’t support gin index - we remove this endpoint, still much much faster - they still blame us, we explain how and why we are faster - they realise there’s nothing they can do with the truth, in their own short tests drizzle latency is 4 times lower - they release their own benchmarks where they claim to be faster, don’t accept prs which make tests fair, don’t accept indexes in schemas, because it would ruin the “drizzle and prisma are about the same performance” completely - their benchmarks is running 500 queries sequentially - they release prisma 7, claim they have x3 perf of queries, claim it’s now typescript - people start adoption, production app break due to performance issues, cold starts in lambdas are bigger - they come out with acknowledgment of a regression, though claim the load is not production like, but of a microbenchmark type. The same type they have in their own benchmarks - nobody knows what tests they run with x3 claim, their own benchmarks do not have a branch with prisma 7. They don’t share which tests they did - drizzle benchmarks are still http server with a db that just receives requests with progressive load increase I can’t believe it’s a database vendor

We're happy to announce the world’s second ever Bitcoin-backed line of credit (BLOC). Why second? Because “first” got announced yesterday. Ours has been running all year and has originated >$350M. Your BTC loan already acts like a BLOC on Arch: • Draw more cash when LTV < 50% • Repay anytime with zero prepayment fees • Rollover indefinitely • USDC or USD funded in a few minutes Silver never looked so good. 🥈

How is "Our status page provider is currently down" still a valid excuse?

Appreciate the feedback. Are you referring to our marketing page calculator or in app? We do currently display all the thresholds and prices on the in-app application but it is on the second screen (so not ideal). We're working on designs for a more holistic in app calculator covering both borrowing needs and health results.

@ArchLending Just an idea for your homepage you should add in the calculator based on what someone wants to borrow what price bitcoin would need to fall too in order to be liquidated from loan.

The smartest Bitcoin play isn't "buy and hold." It's "borrow and hold more." Saylor proves it: Borrow from you at 10.5% with $STRC, buy BTC appreciating 50%+, pocket 40% spread. You can do this too: Borrow against BTC at less than 12% → Buy more Bitcoin → Capture ~38% arbitrage. That's how conviction compounds.

Very interesting to find the ones that are completely taken across all TLDs.

It is incredible how effective @Fin_ai is. @eoghan and the @intercom team really nailed it. All you need to do is keep your help center up to date. We've been able to scale our business in a way that was not possible 2 years ago.